Conformational Changes in the 5-HT3A Receptor Extracellular Domain Measured by Voltage

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

1. Title page

Clamp Fluorometry

Lachlan Munro1, Lucy Kate Ladefoged2, Vithushan Padmanathan1, Signe Andersen1, Birgit

Schiøtt2, 3, Anders S. Kristensen1

Downloaded from

1Department of Drug Design and Pharmacology, University of Copenhagen, Universitetsparken 2,

DK-2100 Copenhagen, Denmark molpharm.aspetjournals.org 2 Department of Chemistry, Aarhus University, Langelandsgade 140, DK-8000 Aarhus, Denmark

3Interdisciplinary Nanoscience Center (iNANO), Aarhus University, Langelandsgade 140, DK-

8000 Aarhus, Denmark at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

2. Running title page

Running title:

Ligand Specific Conformational Changes at the 5-HT3 Receptor

Corresponding author information:

Anders S. Kristensen, Department of Drug Design and Pharmacology, University of Copenhagen,

Universitetsparken 2, DK-2100 Copenhagen, Denmark. E-mail: [email protected]. Tel.: +45 35 30 Downloaded from

65 05

molpharm.aspetjournals.org Abbreviations used:

5-HT: 5-hydroxytryptamine; 5-HTBP: 5-HT binding protein; ACh: Acetylcholine; AChBP:

Acetylcholine binding protein; CNS: central nervous system; Cryo-EM: Cryo-electron microscopy;

DMEM: Dulbecco's modified Eagle's medium; ECD: extracellular domain; FPS: FRET positioning at ASPET Journals on October 1, 2021 and screening; GABA: γ-aminobutyric acid; h5-HT3A: human 5-HT3A; m5-HT3A: mouse 5-HT3A; mCPBG: 1-(m-chlorophenyl)-biguanide, MD: molecular dynamics; MTS-TAMRA: 2-((5(6)- tetramethylrhodamine)carboxylamino)ethyl methanethiosulfonate; pLGIC: pentameric ligand-gated ion channel; PMT: photomultiplier tube; TEVC: two-electrode voltage clamp; TMD: transmembrane domain; VCF: voltage-clamp fluorometry.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

3. Abstract

The 5-HT3 receptor is a member of the Cys-loop receptor super family of ligand-gated ion channel in the nervous system and is a clinical target in a range of diseases. The 5-HT3 receptor mediates fast serotonergic neurotransmission by undergoing a series of conformational changes initiated by ligand-binging that lead to the rapid opening of an intrinsic cation-selective channel.

However, despite the availability of high-resolution structures of a mouse 5-HT3 receptor, many important aspects of the mechanistic basis of 5-HT3 receptor function and modulation by drugs Downloaded from remain poorly understood. In particular, there is little direct evidence for the specific conformational changes predicted to occur during ligand-gated channel activation and molpharm.aspetjournals.org desensitization.

In the present study, we used voltage-clamp fluorometry (VCF) to measure conformational changes in regions surrounding the orthosteric binding site of the human 5-HT3A (h5-HT3A) receptor during binding of 5-HT and different classes of 5-HT3 receptor ligands. VCF utilizes parallel at ASPET Journals on October 1, 2021 measurements of receptor currents with photon emission from fluorescent reporter groups covalently attached to specific positions in the receptor structure. Reporter groups that are highly sensitive to the local molecular environment can in real-time report conformational changes as changes in fluorescence that can be correlated with changes in receptor currents reporting the functional states of the channel. Within the loop C-, D-, and E-regions that surround the orthosteric binding site in the h5-HT3A receptor, we identify positions that are amenable to tagging with an environmentally sensitive reporter group that reports robust fluorescence changes upon 5-HT binding and receptor activation. We use these reporter positions to characterize the effect of ligand binding on the local structure of the orthosteric binding site by agonists, competitive antagonists, and allosterically acting channel activators. We observed that loop C appears to show distinct fluorescence changes for ligands of the same class, while loop D reports similar fluorescence

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

changes for all ligands binding at the orthosteric site. In contrast, the loop E reporter position shows distinct changes for agonists, antagonists, and allosteric compounds, suggesting the conformational changes in this region are specific to ligand function. Interpretation of these results within the framework of current models of 5-HT3 and Cys-loop mechanisms are used to expand the understanding of how ligand binding in Cys-loop receptors relates to channel gating. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

4. Significance statement

The 5HT3 receptor is an important ligand-gated ion channel and drug target in the central and peripheral nervous system. Determining how ligand binding induced conformational changes in the receptor is central for understanding the structural mechanisms underlying 5HT3 receptor function.

Here, we employ voltage-gated fluorometry to characterize conformational changes in the extracellular domain of the human 5HT3 receptor to identify intra-receptor motions during binding of a range of 5HT3 receptor agonists and antagonists. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

5. Introduction

5-HT3 receptors in the central nervous system transduce fast synaptic transmission and are considered therapeutic targets for the treatment of chemotherapy-induced nausea and vomiting and psychiatric conditions such as schizophrenia and depression (Thompson & Lummis, 2007; Walstab et al., 2010). Furthermore, 5-HT3 receptors in the enteric nervous system regulate gut motility and are well-established targets for the treatment of irritable bowel syndrome (Lummis, 2012). 5-HT3 receptors belong to the pentameric ligand-gated ion channel (pLGIC) superfamily, also known as Downloaded from

Cys-loop receptors (Nemecz et al., 2016). pLGICs have as structural hallmark five subunits forming a central ion-permeable pore in the transmembrane domain (TMD; Fig. 1A) with orthosteric ligand molpharm.aspetjournals.org binding sites located at subunit interfaces in the extracellular domain (ECD; Fig. 1A). pLGICs are thought to share a common structural mechanism for how agonist binding in the ECD couples to channel activation and desensitization in the TMD (Grosman et al., 2000; Barnes et al., 2009;

Corringer et al., 2010; Lummis, 2012; Nys et al., 2013; Keramidas and Lynch, 2013; Nemecz et al., at ASPET Journals on October 1, 2021 2016).

Detailed structural information is available for the 5-HT3 receptor in the form of X-ray crystal and cryo-electron microscopy (cryo-EM) structures of the mouse 5-HT3A (m5-HT3A) receptor

(Hassaine et al., 2014; Basak, Gicheru, Rao, et al., 2018; Basak, Gicheru, Samanta, et al., 2018;

Polovinkin et al., 2018). Also, the acetylcholine binding protein (AChBP), which is a soluble pentameric eukaryotic protein homologous to the ECD of Cys-loop receptors (Brejc et al., 2001;

Sixma & Smit, 2003) has been engineered to mimic the 5-HT3 receptor ligand binding profile

(Kesters et al., 2012). Crystal structures of this construct (5-HT binding protein, denoted 5-HTBP) in complex with 5-HT, granisetron (Kesters et al., 2012), palonosetron (Price et al., 2016) and varenicline (Price et al., 2015) provide atomic-level insight to potential ligand binding modes and local conformations of the ECD in the 5HT3 receptor. However, the binding mode of 5-HT, as well

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

as the local conformation of the binding sites observed in the 5-HTBP structure (Kesters et al.,

2012), is different to the binding modes of 5-HT observed in recently reported cryo-EM structures of the mouse 5-HT3 (m5-HT3) receptor (Basak et al., 2018a; Polovinkin et al., 2018). Thus, there are still gaps in the understanding of how ligand binding in the orthosteric site can either induce or prevent the opening of the relatively distal ion channel; partly due to a lack of data describing specific intra-receptor motions during ligand binding.

Voltage-clamp fluorometry (VCF) is a technique that uses site-specific labeling of proteins with Downloaded from an environmentally sensitive fluorophore coupled with two-electrode voltage-clamp electrophysiology (TEVC) (Mannuzzu et al., 1996). VCF allows simultaneous real-time molpharm.aspetjournals.org measurement of both the functional state of an ion channel and conformational changes around specific regions, which are reported as changes in fluorescence intensity (∆F) (Talwar & Lynch,

2015). Previously, VCF has been used to investigate conformational changes in the ECD of other

Cys-loop receptors such as the glycine receptor (Pless & Lynch, 2009a; Lynagh & Lynch, 2012; at ASPET Journals on October 1, 2021

Soh et al., 2017), the GABAA receptor (Chang & Weiss, 2002; Khatri et al., 2009; Muroi et al.,

2009; Akk et al., 2011; Eaton et al., 2014), the nicotinic acetylcholine receptor (Dahan et al., 2004) and very recently, used to track conformational changes in the ion pore region of the m5-HT3 receptor associated with channel gating (Polovinkin et al., 2018).

In the present study, we utilize VCF to monitor conformational changes in the ECD around the orthosteric site of the human 5-HT3A (h5-HT3A) receptor induced by binding of agonists, antagonists, and allosteric modulators. At the 5-HT3 subunit interfaces, the orthosteric binding sites are formed by six segments (designated loop A to F; Fig 1). Specifically, we identify reporter positions in loop C, E and F where residues are amenable to replacement with cysteine (Cys) and labeling with the fluorophore 2-([5(6)-tetramethyl-rhodamine] carboxylamino) ethyl methanethiosulfonate (MTS-TAMRA) and can report ligand-induced structural rearrangements as

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

changes in TAMRA fluorescence. In agreement with VCF studies of other Cys-loop receptors, our findings indicate that the loop E reporter position display distinct ∆F for agonists and antagonists. A loop D reporter position was found to report similar ∆F in response to all orthosteric binding ligands, regardless of function. We also observe that functionally and structurally similar antagonists can induce distinct ∆F at a reporter position in loop C. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

6. Materials and Methods

Materials – Chemicals were from Sigma (St. Louis, MO) except when otherwise stated. DNA modifying enzymes were from New England Biolabs (Ipswich, MA). Granisetron, ondansetron, tropisetron, mirtazapine, thymol, and carvacrol were from Tokyo Chemical Industry (Zwijndrecht,

Belgium). m-chlorophenyl biguanide (mCPBG) was from Tocris Bioscience (Bristol, UK). 2-([5(6)-

Tetramethyl-rhodamine] carboxylamino) ethyl methanethiosulfonate (MTS-TAMRA) was from

Biotium (Hayward, CA). Downloaded from

Molecular Biology – cDNA encoding h5-HT3A was kindly provided by Beate Niesler (Dept. of Human Molecular Genetics, University of Heidelberg, Heidelberg). For expression in Xenopus molpharm.aspetjournals.org laevis oocytes, the h5-HT3A coding region was subcloned into the pXOON expression vector as described previously (Ladefoged et al., 2018). Cysteine (Cys) mutations were introduced using the

QuickChange site-directed mutagenesis kit from Stratagene (La Jolla, CA). Successful mutagenesis

was confirmed by sequencing of the entire coding region (GATC Biotech, Constance, Germany). at ASPET Journals on October 1, 2021 mRNA was generated using the AmpliCap T7 High Yield Message Maker Kit (CellScript,

Madison, WI) according to the manufacturer’s instructions.

Xenopus Laevis Oocyte Expression – Defoliated stage V to VI oocytes from Xenopus laevis were prepared as described previously (Poulsen et al., 2014) and injected with 15 ng mRNA. The care and use of Xenopus laevis were in strict adherence to a protocol (license

2014−15−0201−00031) approved by the Danish Veterinary and Food Administration. Injected oocytes were incubated at 18 °C for 3 to 10 days in Standard Barth’s Medium containing (in mM)

88 NaCl, 1 KCl, 0.41 CaCl2, 2.4 NaHCO3, 0.33 Ca(NO3)2, 0.82 MgSO4, 5 Tris (pH 7.4) supplemented with 50 µg/ml gentamycin. Healthy oocytes that showed robust expression were stored at 4 °C for up to 5 additional days.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Labeling of oocytes – Oocytes were transferred to frog Ringer’s solution (115 mM NaCl, 2 mM KCl, 5 mM HEPES, 1.8 mM BaCl, pH 7.6) containing 20 µM MTS-TAMRA and left for 30 seconds in the dark at room temperature. Oocytes were then rinsed three times in frog Ringer’s solution. Labeled oocytes were kept in Ringer’s solution in the dark for up to two hours before recording.

Electrophysiology and Voltage Clamp Fluorometry – VCF recordings were performed essentially as described previously (Soderhielm et al., 2015). A glass bottom perfusion chamber was Downloaded from mounted on an inverted epifluorescence microscope (Diaphot 300; Nikon, Japan) equipped with a

20x air objective and a G-1B filter set (Chroma 11002v2, excitation filter 541-551 nm, dichroic molpharm.aspetjournals.org mirror > 565 nm, emission > 590 nm). Excitation was achieved using an LED lamp with a nominal wavelength of 530 nm (M530L3; ThorLabs, Newton, NJ). Fluorescence intensity was measured with a D104 photomultiplier tube (PMT) system (Photon Technologies International). PMT voltage

was adjusted daily between 450 and 700 volts such that baseline converted fluorescence output on at ASPET Journals on October 1, 2021 labeled oocytes was between 0.5 and 5 volts. Membrane current was measured under two-electrode voltage clamp (TEVC). Oocytes were voltage-clamped at -20 mV and continuously perfused with frog Ringer’s. All ligands were dissolved in frogs Ringer’s solution and applied using an automated programmable perfusion system (Bioscience Tools, San Diego, CA). Current, voltage and fluorescence signals were digitized at 200 Hz (DigiData 1332, Molecular Devices, Sunnyvale, CA) and recorded using a PC running Clampex 10 software (Molecular Devices, Sunnyvale, CA). For fluorescence antagonist dose-response curves, each concentration was applied for two minutes to allow stable fluorescence levels. The excitation LED was pulsed for 200 ms every second to limit the fluorophore photodestruction during the extended recording time. For fast-solution switching electrophysiology experiments to determine rates of activation, deactivation, and desensitization, we utilized a vertical flow chamber with a chamber volume of 200 µL (Joshi at al., 2004). With a

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

constant flow rate of 20 mL/min, this chamber design results in complete solution exchange within

∼500 ms as measured by recordings open-tip pipette potential when the pipette is positioned just above the oocyte in response to exchange of high and low molarity.

Molecular modeling – A previously generated homology model of h5-HT3A (Ladefoged et al.,

2018) was used as a starting point for the creation of models of h5-HT3A receptor Cys-mutants with

TAMRA conjugated. Specifically, for each reporter position (Tyr89, Met223, and Gln146), the wild-type residue was mutated to a Cys using the mutate tool within Maestro 10.1 (Schrödinger Downloaded from

Suite, 2015). The TAMRA molecule including the linker was built in Maestro 10.1 and minimized using a conjugate-gradient algorithm in MacroModel 10.7. The protonation state of the TAMRA molpharm.aspetjournals.org molecule was assessed using Epik 3.1 (Shelley et al., 2007; Greenwood et al., 2010) and was modeled in a zwitterionic state. TAMRA structural conformation was optimized using MacroModel

10.7 and a 1000-step mixed torsional/low-mode sampling conformational search algorithm in which

a 2500-step minimization followed each step with all other settings at default. Insertion of TAMRA at ASPET Journals on October 1, 2021 and MTS linker at Cys residues was performed by structurally aligning S-Cβ-Cα atoms from the Cys side chain to S-CH2-CH3 heavy atoms from the linker. Duplicate atoms were deleted, and the atomic bond between the linker and the reporter position side chain was drawn manually in Maestro

10.1. The QuickTorsion tool available in Maestro was used to move linker and TAMRA molecule out of direct steric clashes with the protein. The resulting TAMRA-linked h5-HT3A structure was further processed using the Protein Preparation Wizard (Sastry et al., 2013). Specifically, the conjugated TAMRA including linker and protein residues within 5 Å were subjected to energy minimization using a conjugate gradient algorithm and a conformational search using the substructure option in MacroModel 10.7, while residues in the 5 to 10 Å distance range were restrained by a force constant of 200 kJ/(mol Å2). All remaining atoms were ignored during the calculation. Calculation of accessible volume for TAMRA at each reporter position was performed

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

using the FRET positioning and screening (FPS) software (Kalinin et al., 2012). This software estimates the volume accessible to the MTS-linked TAMRA molecule at each reporter position using a geometric search algorithm. The fluorophore is approximated by a sphere, and the linker is simply treated as a distance. The search algorithm determines all possible orientations and positions the fluorophore can reach based on the applied linker that does not clash sterically with the protein.

Before each calculation, the side chain of the reporter positions in subunit A (Met223 from the principal face) or E (Tyr89 and Gln146 from the complementary face) was removed to prevent the Downloaded from original residue from causing steric clashes with the fluorophore probe. Three parameters were used to describe MTS-TAMRA during calculations: linker length (16 Å), linker width (3 Å), and dye molpharm.aspetjournals.org radii (x, y, z is 7, 5, and 3 Å, respectively). The dimensions were estimated based on a computational model of MTS-TAMRA in a stretched conformation, and the linker length is the distance from the distal end of the linker to the geometric center of TAMRA.

Data and Statistical Analysis –Data and statistical analyses were performed using pClamp 10 at ASPET Journals on October 1, 2021

(Molecular Devices) or Prism 6.0 (GraphPad Inc., San Diego, CA) software. Calculation of ∆F was performed using the equation:

Where Fpeak is the peak fluorescence observed following application of ligand and Fbaseline is the fluorescence immediately before ligand application. For EC50 determination, current or fluorescence responses were normalized to the maximal ∆F or current response in each oocyte and pooled. The composite concentration-response data analyzed by iterative curve fitting using the equation:

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

where Response is the ∆F or current response measured at a given concentration normalized to the maximum response, Y is the concentration of ligand that produces a half-maximal response (EC50),

X is the logarithm of the concentration of the ligand, and nH is the Hill slope. Unless otherwise noted, data are reported as either mean ± SEM or mean with 95 % confidence intervals. Statistical analysis of pairwise or multiple comparisons was performed using ANOVA or Student's t-test as appropriate. A probability value (p) < 0.05 was considered statistically significant. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

7. Results

Identification of VCF reporter positions in h5-HT3A – To investigate ligand-induced structural changes in 5-HT3 receptors, we first searched for residues in homomeric h5-HT3A receptors that were amenable to substitution with Cys and thiol-mediated fluorophore conjugation. However, a prerequisite of this type of site-specific fluorophore labeling is the absence of endogenous Cys residues that are extracellularly accessible and thiol-reactive. Structures of the m5-HT3A receptor

(Hassaine et al., 2014; Basak et al., 2018a; Polovinkin et al., 2018; Basak et al., 2018b; Basak et al., Downloaded from

2019) show that the only Cys residues in the ECD are Cys157 and Cys171 in the eponymous Cys- loop. Cys157 and Cys171 form a disulfide bond and thus should not be reactive towards MTS- molpharm.aspetjournals.org TAMRA (Fig. 1B). Also, previous experiments utilizing thiol-reactive fluorophores on other Cys- loop receptors have found this pair of conserved Cys residues to be inaccessible to labeling (Dahan et al., 2004; Pless & Lynch, 2009a; Eaton et al., 2014). However, to verify that potential labeling of

Cys157 and Cys171 did not produce background signals, we initially subjected oocytes expressing at ASPET Journals on October 1, 2021 wild-type (WT) h5-HT3A receptors to MTS-TAMRA labeling and measured fluorescence during application of different ligands (Materials and methods). Following treatment with MTS-TAMRA, application of 100 µM 5-HT to WT h5-HT3A expressing oocytes showed no or less than 1 % ∆F and was in general indistinguishable from fluorescence recordings at MTS-TAMRA treated uninjected oocytes (Fig. 1C and D). On this basis, we concluded that the WT h5-HT3A receptor formed a suitable background for the introduction of Cys residues for VCF.

We next sought to identify residues in the h5-HT3A receptor subunit useful as VCF reporter positions following substitution with Cys and MTS-TAMRA labeling. Criteria for a residue position to be a useful reporter include that i) the position must allow the side chain of the introduced Cys residue to be accessible to an extracellularly applied thiol-reactive fluorophore, ii)

Cys-mutation does not interfere excessively with subunit assembly and normal receptor function,

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

and iii) be capable of reporting ∆F in response to conformational changes at the labeled position.

Extensive cysteine scans often yield only a small number of useful VCF reporter positions (Zhang et al., 2009; Talwar & Lynch, 2015). To improve the rate of identifying successful reporter residues, we used previously performed VCF work on other Cys-loop receptors as a guide to residue selection (Table 1 and Supplementary Fig. S1). On this basis, seven h5-HT3A residues around the orthosteric binding site were selected for mutation to Cys (Fig. 1B and Table 1). Also, we sought to identify reporter positions for movements occurring outside the orthosteric binding site; potentially Downloaded from leading toward channel activation. The pre-M1 region is known to couple conformational changes between the orthosteric binding site and the channel and to rearrange before opening of the channel molpharm.aspetjournals.org gate (Castaldo et al., 2004; Purohit et al., 2013), which via structures of the 5HT3A receptor in various functional states have been identified to be formed by residues located roughly in the middle of transmembrane helix M2 that lines the pore (Polovinkin et al., 2018; Basak, Gicheru,

Rao, et al., 2018). We, therefore, performed scanning Cys mutagenesis along the pre-M1 region; at ASPET Journals on October 1, 2021 introducing Cys at a further seven positions (Fig. 1B and Table 1).

We expressed Cys-mutant receptors individually in Xenopus oocytes and, following labeling with MTS-TAMRA, evaluated for current and ∆F responses to the application of 100 µM 5-HT

(Materials and methods) (Fig. 1B). As summarized in Fig. 1D, all tested mutants showed inward currents in response to 5-HT with amplitudes that were within the range observed for the WT receptor. Four mutants (Y89C, M223C, F221C, and Q146C) were found to report robust ∆F in response to 100 µM 5-HT (Fig. 1C and D). These mutants all showed increased fluorescence in response to 5-HT with mean ∆F values of 30 ± 2 % (Y89C), 11 ± 1 % (Q146C), 4.6 ± 0.7 %

(F221C), and 14 ± 1 % (M223C) (Fig. 1B and C). The remaining mutants did not report detectable

∆F in response to 100 µM 5-HT or produced ∆F that were less than 2-fold of background fluorescence and were not investigated further. A caveat of VCF when using reporter positions near

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

ligand binding sites is that fluorescence changes may theoretically indicate a change in microenvironment due to proximity to the ligand and not a conformational change in the protein.

We determined the effect of 5-HT on TAMRA fluorescence in solution (Materials and Methods). 5-

HT caused a concentration-dependent decrease of TAMRA fluorescence intensity with no change in the peak emission wavelength (Supplementary Fig. S2). Thus, if bound 5-HT exerts a direct effect on the fluorescence of receptor-conjugated TAMRA, this would be reported as a decrease in fluorescence or negative ∆F. As the ∆F observed during VCF at all positions were positive, we Downloaded from concluded that the observed changes in fluorescence are not caused by direct fluorophore-ligand interaction. Consequently, we assume that the ∆F observed in response to 5-HT indicates local or molpharm.aspetjournals.org global conformational changes of the protein.

Characterization of reporter position Cys mutants – To evaluate if Cys-mutation and TAMRA- labeling change basal ligand-gated ion channel properties of the potential reporter Cys-mutants, we performed electrophysiological characterization of 5-HT response kinetics and EC50 (Fig. 2 and at ASPET Journals on October 1, 2021

Table 2). First, we determined EC50 for 5-HT at WT and the Y89C, M223C, Q146C, and F221C mutant receptors with and without MTS-TAMRA treatment (Fig. 2A)..With and without labeling, the Y89C and Q146C mutant showed EC50 values that were within close range of the EC50 for the

WT receptor (Table 2). In contrast, there was a profound decrease in 5-HT potency for the F221C mutant with a 47-fold increase in EC50, which upon TAMRA labeling further increased to more than 150-fold relative to WT (Table 2). The M223C mutant displayed a 3-fold increase in EC50 relative to WT and increased further to 12-fold upon TAMRA labeling (Table 2). The Phe221 and

Met223 positions are located very near each other on loop C and therefore likely report the same conformational changes. The VCF screening results suggested that the M223C mutant, in general, displayed stronger fluorescence signals compared to F221C (Fig. 1C). Therefore, due to the

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

profound effect of Cys-mutation and TAMRA labeling on the EC50 for 5-HT for the F221C mutant and its generally weaker fluorescence signals, we abandoned further work with the F221C mutant.

To further characterize the potential effects of Cys-mutations and TAMRA-labeling on receptor function, we performed fast ligand-application experiments to determine the 5-HT response kinetics. Specifically, using a vertical oocyte perfusion chamber that allows solution exchange rates in the 100 to 500 ms range (Joshi et al., 2004) (Methods and materials), we characterized the rate of activation, desensitization and deactivation of currents evoked by application of saturating Downloaded from concentration (1 mM) of 5-HT to oocytes expressing WT and Y89C, Q146C, and M223C mutant receptors (Fig. 2B). The obtained time constants for activation (τact), deactivation (τdeact), and molpharm.aspetjournals.org desensitization (τdesens) were in good agreement with previously reported characteristics of 5-HT3A receptor currents recorded from oocytes (Jackson & Yakel, 1995; Yakel, 1996; Davies et al., 1999;

Dubin et al., 1999; Joshi et al., 2004). Furthermore, the results showed that the Cys-mutations and

TAMRA labeling has no or little effect on receptor kinetics; except for a modest 3-fold decrease of at ASPET Journals on October 1, 2021 the activation rate for the Q146C mutant upon TAMRA labeling (Fig. 2C). In summary, the Y89C,

Q146C, and M223C mutants overall retain receptor properties similar to WT, and we, therefore, selected all three mutants for further VCF experimentation.

Modeling TAMRA at the three reporter positions - To investigate how the Cys-conjugated

TAMRA molecule at the 89, 146, and 223 positions relate to the orthosteric binding site, we generated models of the h5-HT3A receptor in the inactive conformation with the fluorophore attached at the three reporter sites (Fig. 3A) and calculated the volume accessible to the fluorophore at each position (Materials and Methods). The models show that the fluorophore at the 89 and 146 positions in loop D and E, respectively, are well-accommodated by the protein and can move freely around the complementary face of the binding site (Fig. 3B and C). The accessible volume at these two reporter positions overlap directly in front of loop D and E moreover, are also close to loop C.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

However, only at position 89, TAMRA can reach loop B from the principal face and only at position 146, TAMRA can reach loop F and G (Fig. 3B and C). Due to the flexible nature of the linker and loop C itself, TAMRA at position 223 can sample a large portion of both the principal and complementary subunit when loop C is in an extended conformation such as observed in structures thought to represent the receptor in an inactive state (Fig. 3D). In contrast, when loop C is closed in on the binding site as is suggested to occur during activation, movement of the fluorophore is likely dynamically restricted by loop C as the Cys side chain at position 223 is facing Downloaded from the binding site. Importantly, based on the accessible volume calculations, the TAMRA molecule at position 223 is able to occupy the orthosteric binding site and possibly compete with the tested molpharm.aspetjournals.org ligands, which may explain the increase in agonist EC50 observed for the M223C mutant following

TAMRA labeling (Fig. 2A). In summary, the results from the TAMRA modeling indicate that the

TAMRA conjugated at position 89, 146, and 223 collectively cover the space surrounding the orthosteric binding site. at ASPET Journals on October 1, 2021

Fluorometric characterization of ligand-induced conformational changes – At present, there is a complete lack of direct experimental evidence of the conformational changes in the ECD that are induced by different classes of 5-HT3 ligands and their correlation to ligand functional effects. We, therefore, measured ∆F in response to a range of structurally and functionally diverse 5-HT3 ligands

(Fig. 4A) at the Y89C, M223C, and Q146C mutants. These ligands included the orthosteric ligands acting as agonists (5-HT and mCPBG) and competitive antagonists (granisetron, ondansetron, tropisetron, and encenicline) as well as the positive allosteric modulators thymol and carvacrol, which are suggested to bind at a transmembrane site (Lansdell et al., 2014) (Fig. 4A). Figure 4B shows the standard recording protocol used to compare the ∆F induced by different ligands. Briefly, the protocol employs an initial application of 100 µM 5-HT followed by a washout period to allow full recovery from desensitization of the receptors before the application of test ligand. Controls for

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

non-specific effects of ligands on TAMRA fluorescence included experiments on oocytes expressing WT h5-HT3A or uninjected oocytes subjected to similar MTS-TAMRA labeling protocols (Supplementary Fig. S3). These showed that saturating concentrations of all ligands evoked less than 1 % ∆F at labeled uninjected or WT h5-HT3A expressing oocytes, except for thymol and carvacrol which at 1 mM caused decreasing ∆F in the -1 to -3% range at both uninjected and WT expressing oocytes. However, these differences were considered small relative to the 5-HT induced fluorescence changes at the reporter positions, and thus thymol and carvacrol were Downloaded from characterized for ∆F at the reporter positions.

Ligand-specific fluorescence signals for the loop D mutant Y89C – Tyr89 is located on loop D, molpharm.aspetjournals.org which is part of a β-sheet motif forming the back wall of the orthosteric binding site (Fig. 3). For this reporter position, all tested ligands displayed ∆F responses to application of saturating concentrations that were very similar in amplitude and direction as the ∆F observed for 5-HT;

except for thymol and carvacrol, which did not evoke detectable specific ∆F (Fig. 4C). Figure 5A at ASPET Journals on October 1, 2021 summarizes the amplitude pattern of the ligand-specific ∆F responses normalized to the response amplitude of the ∆F to evoked by saturating concentration of 5-HT. The orthosteric ligands mCPBG, granisetron, ondansetron, tropisetron, and encenicline all induced ∆F with amplitudes that were not significantly different from those induced by 5-HT (Fig. 5A). These results indicate that the local environment surrounding loop D may experience similar changes in physiochemical properties in response to binding of orthosteric agonists and competitive antagonists.

Ligand-specific fluorescence signals for the loop E mutant Q146C – Gln146 is located on loop

E, which like loop D forms part of the β-sheet at the back of the binding site (Fig. 3). All ligands tested at the TAMRA-labelled Q146C mutant caused robust increases in TAMRA fluorescence upon application, except thymol and carvacrol that did not induce specific ∆F (Fig. 5B). 5-HT and the synthetic agonist mCPBG induced similar ∆F values, suggesting that both agonists induce

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

similar local conformational changes around loop E upon binding and receptor activation. In contrast, all antagonists induced ∆F that was significantly larger than 5-HT (300 to 600%; Fig. 4B and 5B). Q146C, therefore, appears to report the regions around loop E to undergo conformational changes upon orthosteric ligand binding that are distinct for agonists and antagonists.

Ligand-specific fluorescence signals for the loop C mutant M223C – Met223 is located on loop

C (Fig. 3), a region on the principal subunit which is thought to be flexible and adopt a contracted conformation upon agonist binding and an extended conformation when antagonists are bound Downloaded from

(Kesters et al., 2012; Polovinkin et al., 2018). At TAMRA-labeled M223C mutant 5-HT3A receptors, all tested ligands reported an ∆F in the same direction as 5-HT (Fig. 4E). However, in molpharm.aspetjournals.org contrast to the relative ∆F patterns observed at the Y89C and Q146C mutant receptors, there was no clear trend in terms of the relative amplitude of the ∆F values for agonist or antagonist or positive allosteric modulators (Fig. 5C). mCPBG displayed ∆F values that were 34% of those for 5-HT (Fig.

5C). For the four antagonists, ondansetron and encenicline induced ∆F that were 70 to 130% larger at ASPET Journals on October 1, 2021 than 5-HT, respectively, whereas granisetron and tropisetron displayed similar ∆F compared to the

5-HT application. Thymol and carvacrol evoked currents of similar amplitude as 5-HT, and both caused a small, but significant ∆F with amplitudes of 22% and 15%, respectively, of the 5-HT evoked ∆F (Fig. 4E and 5C).

Concentration-response relationship of ligand-specific fluorescence signals - The robust ligand- evoked ∆F responses for Y89C, Q146C, and M223C allowed the determination of steady-state concentration-response relationships for the fluorescence responses (Fig. 6, Table 2). Importantly, this enabled us to determine the EC50 of the ∆F response for the electrically silent process of antagonist binding (Table 3). At Y89C receptors for the endogenous agonist 5-HT, we found that the fluorescence EC50 was increased by approximately 6-fold compared to the EC50 obtained from the simultaneously determined current responses (Table 2). A similar result was obtained for the

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

agonist mCPBG, which displayed a 4-fold increase in the EC50 for ∆F responses compared to the

EC50 obtained from current responses (Table 2). M223C receptors reported similar EC50 values for current and ∆F in response to 5-HT, while mCPBG exhibited an EC50 of ∆F responses that were approximately 2.5-fold larger than current responses (Table 2). Q146C receptors showed the largest difference between fluorescence and current responses with 5-HT and mCPBG showing EC50 values for fluorescence responses that were 12- and 45-fold larger than the EC50 values for current responses, respectively (Table 2). The concentration-response relationship for ∆F signals induced Downloaded from by the competitive antagonists granisetron, ondansetron, tropisetron, and encenicline showed ∆F response amplitudes to be concentration-dependent (Fig. 6). In general, the obtained EC50 values for molpharm.aspetjournals.org ∆F were within 10-fold range of previously reported IC50 or Kd values for all of the antagonists

(Hope et al., 1996; Lankiewicz et al., 1998; Ladefoged et al., 2018); except for the Q146C mutant that displayed decreased EC50 values for ∆F compared to Y89C and M223C mutants, which suggest that Cys-mutation and/or TAMRA conjugation decreases ligand affinity. In general, we did not at ASPET Journals on October 1, 2021 attempt to quantify the rate by which the fluorescence signal decayed back towards baseline during the ligand post-application wash-out phase due to the design of our VCF recording chamber which did not allow for solution exchange rates likely required to isolate the rate of ensemble receptor transitions upon removal of extracellular ligand. However, it is interesting to note that the fluorescence response at Q146C for ligands such as encenicline and ondansetron consistently decreased much faster during the wash-out phase compared to the responses at Y89C and M223C

(Fig. 4C-D); indicating that the ligand dissociation rate at TAMRA-labeled Q146C may be increased.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

8. Discussion

Crystal and cryo-EM structures of the m5-HT3A receptor (Hassaine et al., 2014; Basak et al.,

2018a; Polovinkin et al., 2018; Basak et al., 2018b; Basak et al., 2019) and other Cys-loop receptors

(Du et al., 2015; Morales-Perez et al., 2016; Phulera et al., 2018) have established structural conformations that likely reflect key functional receptor states, and thereby provide an essential framework for beginning to understand the structural mechanism of the 5-HT3 receptor. However, as static structures, these do not provide direct evidence for specific protein motions that occur Downloaded from during receptor state transitions. In our present study, we have used VCF to study conformational changes in the h5-HT3A receptor to provide direct experimental evidence of structural changes molpharm.aspetjournals.org around the orthosteric ligand binding site of the h5-HT3A receptor during ligand binding and channel activation. We identify positions in loop C, D, and E where measurement of fluorescence from a conjugated environmentally sensitive fluorophore allow tracking of conformational changes in response to binding of a range of prototypical 5-HT3 ligands; including the endogenous agonist at ASPET Journals on October 1, 2021 5-HT. Interpretation of our fluorescence data in the context of presently available Cys-loop receptor structures presents an opportunity to evaluate and expand current structure-based models for the early steps in the 5-HT3 receptor activation mechanism. Specifically, the comparison of the patterns of fluorescence changes induced by different ligands provides insight into the differences in conformational changes that underlie agonist and antagonist ligand function. However, a potential concern for interpreting changes in fluorescence from the reporter fluorophore as the results of protein rearrangements is whether the fluorophore microenvironment is affected directly by the proximity of the bound ligands. In terms of direct ligand effects that produce decreases in fluorescence, examples include resonance energy transfer to the ligand or simple absorbance of emitted light by the ligand (Unruh et al., 2005; Pless & Lynch, 2009b; Talwar & Lynch, 2015).

However, none of the ligands induce decreases in TAMRA fluorescence (Fig. 4C and 5), and such

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

direct quenching effects are therefore not a concern in the present study. Direct ligand effects that increase fluorescence require that the presence of the ligand changes the physicochemical microenvironment around the fluorophore such that its quantum yield is increased. In the case of

TAMRA, this most likely means increasing overall hydrophobicity; for example, by shielding the fluorophore from polar protein structure elements or water molecules or by direct steric effects that promote the fluorophore to orient towards a more hydrophobic environment. In this context, we observe that our modeling of TAMRA conjugated to the Cys-mutated Tyr89, and Gln146 positions Downloaded from in the h5-HT3A receptor show that the fluorophore in these positions cannot access the binding site, and thus not be in direct contact with ligands (Fig. 3B and C). In contrast, for the Met223 position, molpharm.aspetjournals.org the side chain of the inserted Cys residue is facing into the binding pocket (Fig. 3D), and the accessible volume calculations show that TAMRA from this position can at least partially occupy the binding pocket and thus potentially experience direct effects of orthosteric ligands; an aspect that therefore must be taken into concern for interpretation of the ΔF reported from this position. at ASPET Journals on October 1, 2021 Another important issue for interpreting fluorescence data is that we utilize a steady-state recording methodology on a time-scale that does not allow us to resolve receptor transitions (e.g. apo to pre- active, active to desensitized); meaning that we cannot know whether a change in fluorescence is due to a local conformational change, such as a localized rearrangement of the orthosteric binding site upon ligand binding, or global conformational changes underlying transitions between receptor functional states.

At positions 89 and 146 in the D and the E loop, respectively, the conjugated TAMRA consistently showed increases in fluorescence upon binding of all the tested orthosteric ligands; thus, showing that conformational changes occur around these positions upon ligand binding.

Principally, these changes can involve the local rearrangement of loop D and E, as well as the surrounding protein regions. Regarding the possibility of direct loop movements as a source for ΔF,

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

then both loop E and D, together with loop G, form strands in a β-sheet structure that lines the back of the orthosteric binding site. This β-sheet motif is considered to be rigid, and comparison of its conformation in recent structures of m5-HT3A and other Cys-loop receptors suggest it to undergo little or no overall structural rearrangement during receptor state transitions (Nys et al., 2013;

Hassaine et al., 2014; Du et al., 2015; Basak et al., 2018a; Polovinkin et al., 2018; Basak et al.,

2018b; Basak et al., 2019). Thus, it appears more likely that TAMRA at both positions reports movement of surrounding structural elements. Candidate elements that might influence Downloaded from fluorescence from TAMRA conjugated at loop E and D first and foremost include the C loop, which is positioned in close vicinity of both positions (Fig. 3). This is based on long-standing molpharm.aspetjournals.org models for Cys-loop receptor function that involves the idea that ligand binding induces local movement of loop C around the orthosteric binding site (Chang & Weiss, 2002; Law et al., 2005;

Pless & Lynch, 2009b; Kesters et al., 2012; Puruhit & Auerbach, 2013; Du et al., 2015). However, for the reporter position 89, potential structural rearrangement of loop F is also a possible candidate at ASPET Journals on October 1, 2021 and this loop region contains several residues that are suggested to be directly involved in agonist recognition (Thompson et al., 2006) and assumes different conformations among the presently available 5-HT3 receptor structures (Hassaine et al., 2014; Basak et al., 2018a; Polovinkin et al.,

2018; Basak et al., 2018b; Basak et al., 2019). Likewise, the reporter position 146 may be influenced by movements of loop B. Interestingly, although the N-terminal segment of loop B is locked in a β-strand motif and likely is overall rigid during state transitions, the C-terminal region is not and may be able to adapt locally to ligand binding (Fig. 1A). Overall, this idea is supported by comparison of structures of apo and 5-HT bound 5-HT3A which show the ECD to adopt a more condensed conformation with 5-HT bound; most pronounced around the lower half of the ECD and loops B and C (Polovinkin et al., 2018; Basak et al., 2018b). In addition, a recent structure of 5-

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

HT3A in complex with the antagonist granisetron also shows the ECD to condense when granisetron is bound; including a shift in the positions of loop B and C (Basak et al., 2019).

In contrast to the distinctly different levels of ligand-induced ΔF for agonist and antagonists observed at the Q146C mutant, the Y89C mutant reported strikingly similar levels of ΔF for both agonists and competitive antagonists. This potentially means that if TAMRA conjugated to position

89 in loop D senses ligand-induced conformational changes, then these are distinct from those reported by TAMRA at position 146; despite the two reporter positions being closely located in the Downloaded from binding site (Fig. 3A). However, it should be noted that it is possible that different rearrangement upon agonist and antagonist binding can produce similar net effects on the environment around molpharm.aspetjournals.org TAMRA and thus similar changes in fluorescence. Also, ligand binding may induce changes to the environment around TAMRA that are not directly related to conformational changes. Notably, albeit the accessible volume calculations show substantial overlap in the parts of the receptor

surface that TAMRA can sample for both positions, only TAMRA at position 89 can sample the at ASPET Journals on October 1, 2021 surface near loop D in the back of the pocket (Fig. 3B and 3C). Instead of conformational changes, a possible source for the similar fluorescence response by the Y89C mutant to agonists and antagonists can be that ligand occupation of the binding pocket, in general, displaces one or more water molecules to increase general hydrophobicity and enhance TAMRA fluorescence.

Comparison of the ΔF observed at the Q146C mutant for agonists with those observed for antagonists shows a striking pattern where the agonists 5-HT and mCPBG induce similar ΔF, whereas the four competitive antagonists induce a distinct set of ΔF of likewise similar magnitude

(Fig. 5B). VCF studies on other Cys-loop receptors have labeled positions in loop E and compared the effects of agonists and antagonists (Muroi et al., 2006; Pless & Lynch, 2009b; Akk et al., 2011).

In GABAA receptors, Akk et al. (2011) observed that a fluorophore-labeled loop E position, corresponding to Gln146 in h5-HT3A, reported increases in fluorescence that were several-fold

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

larger in amplitude for the competitive antagonist gabazine than for the orthosteric agonists GABA and muscimol. Also, at the GABAA receptor, Chang et al. (2002) found that loop E position corresponding to position Gln146 in h5-HT3A reported distinct fluorescence changes for GABA and the competitive antagonist 3-APMPA. Similarly, in glycine α1 receptors, Pless et al. (2009) reported that a loop E reporter position, which corresponds to Gln146 in h5-HT3A receptors

(Supplementary Fig S1; Table 1), showed a greater increase in fluorescence in response to the competitive antagonist strychnine than to the agonist glycine. Taken together with the results from Downloaded from the present study, which utilized a wider range of competitive antagonists, loop E of Cys-loop receptors seems to consistently report VCF results which reflect the functional properties of the molpharm.aspetjournals.org ligand. Distinct conformational changes around loop E may thus be a hallmark feature of Cys-loop receptor agonism versus antagonism.

Met223 is located in loop C, and changes in fluorescence for TAMRA conjugated at this position

might be ideal for tracking of movement of loop C during ligand binding as well as the ECD at ASPET Journals on October 1, 2021 tightening occurring during receptor activation; due to its close proximity to the interface of the principal and the complementary subunit of each binding site. At this position, we did observe fluorescence changes for all ligands, but these displayed a range of relative ΔF magnitudes that did not follow any pattern correlated to ligand functionality (Fig. 5C). Movement of Loop C has been proposed a central role in mechanistic models for Cys-loop receptor function. Specifically, binding of ligands that act as agonists have been proposed to contract loop C around the orthosteric binding site, while binding of ligands that act as antagonist do not induce constriction or induce a more extended orientation compared to the apo state (Celie et al., 2004; Hansen et al., 2005; Brams et al.,

2011, Kesters et al., 2012; Alix et al., 2016). This “capping” movement by loop C has been proposed essential for promoting structural transitions towards the opening of the channel gate; either via a global transition or a sequence of local conformational changes of neighboring domains

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

that eventually reach the channel gate. However, another model is that the varying loop C conformations in ligand-bound structures reflect the role of loop C residues in forming contacts with the ligand (Purohit & Auerbach, 2013). In this model, loop C “capping” correlates with activation not by inducing conformational change, but by local stabilization of the binding site in a conformation with high open probability. For the 5HT3A receptor, these ideas can now be evaluated by a series of structures of the m5-HT3A receptor in the apo state and in complex with 5-HT

(Polovinkin et al., 2018; Basak et al., 2018b) and the antagonist tropisetron and granisetron (Basak Downloaded from et al., 2019). Comparison of the orientation of the C loop between apo and 5-HT bound structures indeed show that the C loop contracts around the orthosteric binding; moving towards loop E and molpharm.aspetjournals.org D, whereas the orientation of loop C in the tropisetron-bound structure closely resembles the apo orientation; hereby supporting the model with loop C contraction as a key conformational change for receptor activation. However, the granisetron-bound structure also shows contraction of loop C around the binding site that is similar in direction with the 5-HT bound structure; although 5-HT at ASPET Journals on October 1, 2021 appears to induce most contraction (Basak et al., 2019). We find binding of 5-HT, granisetron, and tropisetron to all produce increases in fluorescence from TAMRA conjugated to position 223 in loop C; with 5-HT and granisetron producing identical responses that are slightly higher in magnitude than tropisetron (Fig. 5B). Overall, these results suggest that loop C is closing around these ligands, which potentially indicate a more complex role of loop C movement for receptor activation or inhibition than suggested by the aforementioned existing models. However, as discussed previously, a significant concern is that TAMRA at position 223 may occupy the binding pocket and thus experience direct effects from ligand binding. Thus, further work is needed to explore other loop C positions that do not allow the fluorophore to potentially occupy the binding pocket while serving as reporters for loop C movement in the 5-HT3 receptor as well as the arrival of structures of 5-HT3A in complex with additional agonists and antagonists.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Thymol and carvacrol are structurally closely related positive allosteric modulators of the h5-

HT3A receptor. Their binding site is proposed to be located in the transmembrane domain at the subunit interface (Lansdell et al., 2014), which also has been suggested as the binding site for the positive allosteric modulator TMPPAA (Gasiorek et al., 2016; Polovinkin et al., 2018) not explored in the present study. The M223C mutant showed slight, but significant increases in fluorescence in response to both compounds (Fig. 5). As M223C is located on loop C, which connects directly into the TMD region, this result may demonstrate the allosteric coupling between the orthosteric site and Downloaded from the conformational changes that underlie channel activation as has previously been proposed (Du et al., 2015). molpharm.aspetjournals.org In summary, we have established a VCF approach for studies of conformational changes in the

ECD region of the h5-HT3A receptor that contains the orthosteric binding site. We utilize this to compare the conformational effects of different classes of 5-HT3 receptor ligands to define various

ligand-specific aspects of ligand binding. First, consistent with current predictions from structural at ASPET Journals on October 1, 2021 and biochemical studies, our results support the idea that agonists and competitive antagonists induce distinct conformational changes in the orthosteric binding site. Second, the nature of the ligand-induced conformational changes around loop E appears to be a determinant for ligand functionality as either agonist or antagonist most likely by assessing the inter-subunit tightness in the ECD. Overall, our observations hereby provide insight into the initial steps of the conformational cycle that underlies the ligand-gated ion channel function of the 5-HT3 receptor. We note that we were not successful in identifying reporter positions that could report structural transitions outside the immediate ECD regions containing the orthosteric binding site. VCF work focused on identifying useful reporter positions in the ECD/TMD interfaces is therefore warranted to enable tracking of structural changes further into the receptor functional cycle. Whereas our present work focused on the ECD segment, it is in this respect encouraging to note that recent work

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

on the m5-HT3A receptor has reported a position (Ser296) in the upper region of the pore-lining M2 helix as a useful VCF reporter for structural rearrangements associated with gating transition

(Polovinkin et al., 2018). Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

9. Authorship Contribution

Participated in research design: Munro, Ladefoged, Schiøtt, and Kristensen.

Conducted experiments: Munro, Andersen, Padmanathan, Ladefoged, and Kristensen.

Performed data analysis: Munro, Ladefoged, and Kristensen.

Wrote or contributed to the writing of the manuscript: Munro, Ladefoged, Schiøtt, and Kristensen.

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

10. References

Akk G, Li P, Bracamontes J, Wang M, and Steinbach JH (2011) Pharmacology of structural

changes at the GABAAreceptor transmitter binding site. Br J Pharmacol 162:840–850.

Alix K, Khatri S, Mosier PD, Casterlow S, Yan D, Nyce HL, White MM, Schulte MK, and Dukat

M (2016) Superagonist, Full Agonist, Partial Agonist, and Antagonist Actions of

Arylguanidines at 5-Hydroxytryptamine-3 (5-HT3) Subunit A Receptors. ACS Chem Neurosci

7:1565–1574. Downloaded from

Barnes NM, Hales TG, Lummis SCR, and Peters JA (2009) The 5-HT3receptor - the relationship

between structure and function. Neuropharmacology 56:273–284. molpharm.aspetjournals.org Basak S, Gicheru Y, Kapoor A, Mayer ML, Filizola M and Chakrapani S (2019) Molecular

mechanism of setron-mediated inhibition of full-length 5-HT3A receptor. Nat Commun 10(1):

3225.

Basak S, Gicheru Y, Rao S, Sansom MSP, and Chakrapani S (2018a) Cryo-EM reveals two distinct at ASPET Journals on October 1, 2021

serotonin-bound conformations of full-length 5-HT3A receptor. Nature 563:270–274.

Basak S, Gicheru Y, Samanta A, Molugu SK, Huang W, Fuente ML De, Hughes T, Taylor DJ,

Nieman MT, Moiseenkova-Bell V, and Chakrapani S (2018b) Cryo-EM structure of 5-HT3A

receptor in its resting conformation. Nat Commun 9:514.

Brejc K, Van Dijk WJ, Klaassen R V., Schuurmans M, Van Der Oost J, Smit AB, and Sixma TK

(2001) Crystal structure of an ACh-binding protein reveals the ligand-binding domain of

nicotinic receptors. Nature 411:269–276.

Castaldo P, Stefanoni P, Miceli F, Coppola G, Del Giudice EM, Bellini G, Pascotto A, Trudell JR,

Harrison NL, Annunziato L and Taglialatela M (2004) A novel hyperekplexia-causing

mutation in the pre-transmembrane segment 1 of the human glycine receptor alpha1 subunit

reduces membrane expression and impairs gating by agonists. J Biol Chem 279(24): 25598-

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

25604.

Celie PH, van Rossum-Fikkert SE, van Dijk WJ, Brejc K, Smit AB, Sixma TK (2004) Nicotine and

carbamylcholine binding to nicotinic acetylcholine receptors as studied in AChBP crystal

structures. Neuron 41(6):907-14.

Chang Y, and Weiss DS (2002) Site-specific fluorescence reveals distinct structural changes with

GABA receptor activation and antagonism. Nat Neurosci 5:1163–1168.

Corringer PJ, Baaden M, Bocquet N, Delarue M, Dufresne V, Nury H, Prevost M, and Van Downloaded from

Renterghem C (2010) Atomic structure and dynamics of pentameric ligand-gated ion channels:

New insight from bacterial homologues. J Physiol 588:565–572. molpharm.aspetjournals.org Dahan DS, Dibas MI, Petersson EJ, Auyeung VC, Chanda B, Bezanilla F, Dougherty DA, and

Lester HA (2004) A fluorophore attached to nicotinic acetylcholine receptor beta M2 detects

productive binding of agonist to the alpha delta site. Proc Natl Acad Sci 101:10195–10200.

Davies PA, Pistis M, Hanna MC, Peters JA, Lambert JJ, Hales TG and Kirkness EF (1999) The 5- at ASPET Journals on October 1, 2021 HT3B subunit is a major determinant of serotonin-receptor function. Nature 397(6717): 359-

363.

Du J, Lü W, Wu S, Cheng Y, and Gouaux E (2015) Glycine receptor mechanism elucidated by

electron cryo-microscopy. Nature 526:224–229.

Du J, Dong H and Zhou HX (2012) Size matters in activation/inhibition of ligand-gated ion

channels. Trends Pharmacol Sci 33(9): 482-493.

Dubin AE, Huvar R, D'Andrea MR, Pyati J, Zhu JY, Joy KC, Wilson SJ, Galindo JE, Glass CA,

Luo L, Jackson MR, Lovenberg TW and Erlander MG (1999) The pharmacological and

functional characteristics of the serotonin 5-HT(3A) receptor are specifically modified by a 5-

HT(3B) receptor subunit. J Biol Chem 274(43): 30799-30810.

Eaton MM, Lim Y Bin, Covey DF, and Akk G (2014) Modulation of the human ρ1 GABAA

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

receptor by inhibitory steroids. Psychopharmacology (Berl) 231:3467–3478.

Gasiorek A, Trattnig SM, Ahring PK, Kristiansen U, Frølund B, Frederiksen K, and Jensen AA

(2016) Delineation of the functional properties and the mechanism of action of TMPPAA, an

allosteric agonist and positive allosteric modulator of 5-HT3 receptors. Biochem Pharmacol

110–111:92–108.

Greenwood JR, Calkins D, Sullivan AP, and Shelley JC (2010) Towards the comprehensive, rapid,

and accurate prediction of the favorable tautomeric states of drug-like molecules in aqueous Downloaded from

solution. J Comput Aided Mol Des 24:591–604.

Grosman C, Zhou M, and Auerbach A (2000) Mapping the conformational wave of acetylcholine molpharm.aspetjournals.org receptor channel gating. Nature 403:773–776.

Hansen SB, Sulzenbacher G, Huxford T, Marchot P, Taylor P and Bourne Y (2005) Structures of

Aplysia AChBP complexes with nicotinic agonists and antagonists reveal distinctive binding

interfaces and conformations. EMBO J 24(20): 3635-3646. at ASPET Journals on October 1, 2021 Hassaine G, Deluz C, Grasso L, Wyss R, Tol MB, Hovius R, Graff A, Stahlberg H, Tomizaki T,

Desmyter A, Moreau C, Li X-D, Poitevin F, Vogel H, and Nury H (2014) X-ray structure of

the mouse serotonin 5-HT3 receptor. Nature 512:276–281.

Hope AG, Peters JA, Brown AM, Lambert JJ, and Blackburn TP (1996) Characterization of a

human 5-hydroxytryptamine3receptor type A (h5-HT3R-A(s)) subunit stably expressed in

HEK 293 cells. Br J Pharmacol 118:1237–1245.

Joshi PR, Suryanarayanan A and Schulte MK (2004) A vertical flow chamber for Xenopus oocyte

electrophysiology and automated drug screening. J Neurosci Methods 132(1): 69-79.

Kalinin S, Peulen T, Sindbert S, Rothwell PJ, Berger S, Restle T, Goody RS, Gohlke H, and Seidel

CAM (2012) A toolkit and benchmark study for FRET-restrained high-precision structural

modeling. Nat Methods 9:1218–1225.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Keramidas A1, Lynch JW (2013) An outline of desensitization in pentameric ligand-gated ion

channel receptors. Cell Mol Life Sci 70:1241-53.

Kesters D, Thompson AJ, Brams M, van Elk R, Spurny R, Geitmann M, Villalgordo JM, Guskov

A, Helena Danielson U, Lummis SCR, Smit AB, and Ulens C (2012) Structural basis of ligand

recognition in 5-HT3 receptors. EMBO Rep 14:49–56.

Khatri A, Sedelnikova A, and Weiss DS (2009) Structural rearrangements in loop F of the GABA

receptor signal ligand binding, not channel activation. Biophys J 96:45–55. Downloaded from

Kilpatrick GJ, Butler A, Burridge J, and Oxford AW (1990) 1-(m-Chlorophenyl)-biguanide, a

potent high affinity 5-HT3 receptor agonist. Eur J Pharmacol 182:193–197. molpharm.aspetjournals.org Ladefoged LK, Munro L, Pedersen AJ, Balle T, Bang-Andersen B, Lummis SCR, Schiott B, and

Kristensen AS (2018) Modeling and mutational analysis of the binding mode for the

multimodal antidepressant drug vortioxetine to the human 5-HT3A receptor. Mol Pharmacol

94:1421–1434. at ASPET Journals on October 1, 2021 Lankiewicz S, Lobitz N, Wetzel CHR, Rupprecht R, Unter G, and Hatt H (1998) Molecular

Cloning, Functional Expression, and Receptor cDNA and Its Splice Variants from Guinea Pig.

Mol Pharmacol. 53(2):202-12.

Lansdell SJ, Sathyaprakash C, Doward A, and Millar NS (2014) Activation of Human 5-

Hydroxytryptamine Type 3 Receptors via an Allosteric Transmembrane Site. Mol Pharmacol

87:87–95.

Law RJ, Henchman RH, and McCammon JA (2005) A gating mechanism proposed from a

simulation of a human 7 nicotinic acetylcholine receptor. Proc Natl Acad Sci 102:6813–6818.

Lummis SCR (2012) 5-HT3 receptors. J Biol Chem 287:40239–40245.

Lynagh T, and Lynch JW (2012) Molecular mechanisms of Cys-loop ion channel receptor

modulation by ivermectin. Front Mol Neurosci 5:1–11.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Mannuzzu LM, Moronne MM, and Isacoff EY (1996) Direct physical measure of conformational

rearrangement underlying potassium channel gating. Science 271:213–216.

Menny A, Lefebvre SN, Schmidpeter PA, Drege E, Fourati Z, Delarue M, Edelstein SJ, Nimigean

CM, Joseph D and Corringer PJ (2017) Identification of a pre-active conformation of a

pentameric channel receptor. Elife 15:e23955

Morales-Perez CL, Noviello CM, and Hibbs RE (2016) X-ray structure of the human α4β2 nicotinic

receptor. Nature 538:411–415. Downloaded from

Muroi Y, Czajkowski C, and Jackson MB (2006) Local and global ligand-induced changes in the

structure of the GABA A receptor. Biochemistry 45:7013–7022. molpharm.aspetjournals.org Muroi Y, Theusch CM, Czajkowski C, and Jackson MB (2009) Distinct structural changes in the

GABA A receptor elicited by pentobarbital and GABA. Biophys J 96:499–509.

Mourot A, Bamberg E and Rettinger J (2008) Agonist- and competitive antagonist-induced

movement of loop 5 on the alpha subunit of the neuronal alpha4beta4 nicotinic acetylcholine at ASPET Journals on October 1, 2021 receptor. J Neurochem 105(2): 413-424.

Nemecz Á, Prevost MS, Menny A, and Corringer PJ (2016) Emerging Molecular Mechanisms of

Signal Transduction in Pentameric Ligand-Gated Ion Channels. Neuron 90:452–470.

Nys M, Kesters D, and Ulens C (2013) Structural insights into Cys-loop receptor function and

ligand recognition. Biochem Pharmacol 86:1042–1053.

Phulera S, Zhu H, Yu J, Claxton DP, Yoder N, Yoshioka C, and Gouaux E (2018) Cryo-EM

structure of the benzodiazepine-sensitive α1β1γ2S tri-heteromeric GABAAreceptor in

complex with GABA. Elife 7:1–21.

Pless SA, and Lynch JW (2009a) Distinct conformational changes in activated agonist-bound and

agonist-free glycine receptor subunits. J Neurochem 108:1585–1594.

Pless SA, and Lynch JW (2009b) Ligand-specific conformational changes in the alpha1 glycine

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

receptor ligand-binding domain. J Biol Chem 284:15847–15856.

Polovinkin L, Hassaine G, Perot J, Neumann E, Jensen AA, Lefebvre SN, Corringer PJ, Neyton J,

Chipot C, Dehez F, Schoehn G, and Nury H (2018) Conformational transitions of the serotonin

5-HT3 receptor. Nature 563:275–279.

Poulsen MH, Lucas S, Stromgaard K, and Kristensen AS (2014) Evaluation of PhTX-74 as

Subtype-Selective Inhibitor of GluA2-Containing AMPA Receptors. Mol Pharmacol 85:261-

268. Downloaded from

Price KL, Lillestol RK, Ulens C, and Lummis SCR (2016) Palonosetron-5-HT3 Receptor

Interactions As Shown by a Binding Protein Cocrystal Structure. ACS Chem Neurosci 7:1641– molpharm.aspetjournals.org 1646.

Price KL, Lillestol RK, Ulens C, and Lummis SCR (2015) Varenicline Interactions at the 5-HT 3

Receptor Ligand Binding Site are Revealed by 5-HTBP. ACS Chem Neurosci 6:1151–1157.

Purohit P and Auerbach A (2013) Loop C and the mechanism of acetylcholine receptor-channel at ASPET Journals on October 1, 2021 gating. J Gen Physiol 141(4): 467-478.

Purohit P, Gupta S, Jadey S and Auerbach A (2013) Functional anatomy of an allosteric protein.

Nat Commun 4: 2984.

Sastry GM, Adzhigirey M, Day T, Annabhimoju R, and Sherman W (2013) Protein and ligand

preparation: Parameters, protocols, and influence on virtual screening enrichments. J Comput

Aided Mol Des 27:221–234.

Shelley JC, Cholleti A, Frye LL, Greenwood JR, Timlin MR, and Uchimaya M (2007) Epik: A

software program for pKa prediction and protonation state generation for drug-like molecules.

J Comput Aided Mol Des 12:681–691.

Sixma TK, and Smit AB (2003) Acetylcholine Binding Protein (AChBP): A Secreted Glial Protein

That Provides a High-Resolution Model for the Extracellular Domain of Pentameric Ligand-

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Gated Ion Channels. Annu Rev Biophys Biomol Struct 32:311–334.

Soderhielm PC, Andersen J, Munro L, Nielsen AT, and Kristensen AS (2015) Substrate and

Inhibitor-Specific Conformational Changes in the Human Serotonin Transporter Revealed by

Voltage-Clamp Fluorometry. Mol Pharmacol 88:676–688.

Soh MS, Estrada-Mondragon A, Durisic N, Keramidas A, and Lynch JW (2017) Probing the

Structural Mechanism of Partial Agonism in Glycine Receptors Using the Fluorescent

Artificial Amino Acid, ANAP. ACS Chem Biol 12:805–813. Downloaded from

Talwar S, and Lynch JW (2015) Investigating ion channel conformational changes using voltage

clamp fluorometry. Neuropharmacology 98:3–12. molpharm.aspetjournals.org Thompson AJ, and Lummis SC (2007) The 5-HT 3 receptor as a therapeutic target. Expert Opin

Ther Targets 11:527–540.

Thompson AJ, Padgett CL, and Lummis SCR (2006) Mutagenesis and molecular modeling reveal

the importance of the 5-HT 3 receptor F-loop. J Biol Chem 281:16576–16582. at ASPET Journals on October 1, 2021 Unruh JR, Gokulrangan G, Wilson GS, and Johnson CK (2005) Fluorescence Properties of

Fluorescein, Tetramethylrhodamine, and Texas Red linked to a DNA Aptamer. Photochem

Photobiol 81:682–690.

Walstab J, Rappold G, and Niesler B (2010) 5-HT3 receptors: Role in disease and target of drugs.

Pharmacol Ther 128:146–169.

Zhang J, Xue F, and Chang Y (2009) Agonist- and antagonist-induced conformational changes of

loop F and their contributions to the ρ1 GABA receptor function. J Physiol 587:139–153.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

11. Footnotes

This work was supported by the Danish Council of Independent Research for Medical Sciences

(grant number DFF-404-00309), the Lundbeck Foundation (grant number 2017-1655 and 2012-

12453), and the Carlsberg Foundation. Computations were made possible through allocations at the

Centre for Scientific Computing, Arhus (SCS-Aa). We thank Alexander B. J. Jensen for excellent technical assistance. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

12. Figure legends

Figure 1. Identification of h5-HT3A residues as voltage-clamp fluorometry reporter positions.

A. Surface contoured cartoon representation of a homology model structure of the pentameric h5-

HT3A receptor (left) from Ladefoged et al. (2018). The principal and complementary subunits are highlighted in cyan and green, respectively, with indications of the extracellular domain (ECD) and transmembrane domain (TMD). The cross-sectional view (right) illustrates the location of the five equivalent 5-HT binding pockets (dashed circles) formed at subunit interfaces in the ECD with 5- Downloaded from

HT indicated in yellow surface contour. The chemical structure of the endogenous agonist 5-HT is shown below. B. Zoom of the ECD interface structure in cartoon representation with the orthosteric molpharm.aspetjournals.org binding site and the loop A to G regions highlighted in different colors and bound 5-HT in yellow surface contour. Positions, where Cys residues were introduced by mutagenesis and the endogenous

Cys residues 158 and 171, are indicated as orange and yellow spheres, respectively. C.

Representative recording traces of parallel fluorescence and current measurements from uninjected at ASPET Journals on October 1, 2021 oocytes and oocytes expressing WT h5-HT3A, Y89C, Q146C, and M223C mutants during application of 100 µM 5-HT (black line). Upper traces (red) show the membrane fluorescence signal (F), and lower traces (black) show the corresponding membrane currents (I). D. Summary of fluorescence changes (ΔF) and currents evoked by application of 100 µM 5-HT. Shown are scatter plots of ΔF from individual uninjected oocytes and oocytes injected with WT h5-HT3A or Cys- mutant h5-HT3A. ΔF is calculated as the percentage difference between baseline fluorescence and steady-state fluorescence during 5-HT application (Material and methods). Bars represent mean ΔF

(± S.E.M): uninjected; -0.71 ± 0.28 (13 oocytes), WT; -0.81 ± 0.23% (22 oocytes), D74C; 1.1 ±

0.3% (9 oocytes), Y89C; 30 ± 2% (77 oocytes), Q146C; 11 ± 1 % (80 oocytes), V196C; -1.2 ±

0.4% (9 oocytes), F221C; 4.6 ± 0.7% (48 oocytes), M223C; 14 ± 1% (83 oocytes), S225C; 1.6 ± 0.6

% (7 oocytes), I238C; 0.63 ± 0.15% (8 oocytes), R239C; 1.8 ± 0.4% (8 oocytes), R241C; 0.39 ±

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

0.11% (6 oocytes), L243C; -1.8 ± 0.3% (13 oocytes), V246C; 1.1 ± 0.2% (4 oocytes), V247C; 0.66

± 0.10 (5 oocytes), L249C; 0.56 ± 0.21 (5 oocytes). * Mean response significantly different from

WT; p < 0.05 (ANOVA with Dunnet’s correction for multiple comparisons).

Figure 2. Characterization of effects of Cys-mutation and TAMRA labeling on h5HT3A receptor function. A. Concentration-response curves for 5-HT currents at WT and mutant h5HT3A receptors expressed in oocytes with (red) and without (black) TAMRA labeling (Materials and methods). Data points represent the mean from 3 to 5 independent concentration-response Downloaded from experiments. Error bars are the SEM and are shown when larger than symbol size. B. Superimposed responses to 100-s applications of 1 mM 5-HT (black bar) recorded from oocytes injected with the molpharm.aspetjournals.org indicated WT or Cys-mutant 5-HT3A subunit mRNA; before (black traces) and after TAMRA labeling (red traces). The onset of desensitization (left panels), activation (middle panels) and deactivation (right panels) are fitted with mono-exponential functions (shown as dotted lines) to

determine time constants (τ). Note that for the middle and right panels that show the initial phase of at ASPET Journals on October 1, 2021 the current response, the superimposed responses are shown on a faster time scale compared to left panel and that amplitudes for all responses are scaled to equal sizes. C. Summary of time constants for desensitization (τdesens), activation (τact), and deactivation (τdeact). Data represent the mean and

SEM for 5 to 10 oocytes. * Significantly different from unlabeled WT; p < 0.05 (ANOVA with

Dunnet’s correction for multiple comparisons).

Figure 3. Modeling of the orientation of conjugated fluorophores in the h5-HT3A receptor structure. A. Side view of structural models of TAMRA-conjugation in the mutant Y89C, Q146C, and M223C mutant h5-HT3A receptors. TAMRA is shown in stick representation (red) at position

89 in loop D, position 146 in loop E, and position 223 in loop C (Materials and methods). The

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

agonist 5-HT is shown in yellow surface contour. B-D. The predicted accessible volume for

TAMRA conjugated at Cys residues inserted at position 89 (B), 146 (C), and 223 (D) are shown in gray as observed frontally from within the membrane (top) as well as from the side (bottom) as calculated by the AV method (Materials and methods). The receptor backbone of the principal and complement face of the ECD is shown as light gray ribbons and loops A-F are colored according to the color scheme introduced in Figure 1. Downloaded from Figure 4. Ligand-induced fluorescence changes in fluorophore-labeled h5-HT3A receptors. A.

Chemical structures of 5-HT3 receptor ligands characterized at the Y89C, Q146C, and M223C reporter mutants. B. Representative traces illustrating the standard VCF recording protocol for molpharm.aspetjournals.org measurement of fluorescence (F; upper trace in red) and current (I; lower trace in black) responses to ligand application from oocytes expressing h5HT3A reporter mutants (Materials and methods).

An oocyte expressing the Q146C mutant was exposed 5-HT (black bar; 100 µM) followed by a 6

min wash period (grey bars) before application of a saturating concentration of encenicline (blue at ASPET Journals on October 1, 2021 bar; 10 µM). Excitation light (green bars; 530 nM) was turned off during the wash period to minimize fluorophore photodestruction. Dotted boxes indicate regions of the fluorescence trace that are used for calculation of ΔF for 5-HT and ligand. C-D. Representative parallel fluorescence and current responses to sequential application of 5-HT and ligand (encenicline, 10 µM; ondansetron, 3

µM; granisetron, 3 µM; thymol, 1 mM) at the Y89C (C), Q146C (D), and M223C (E) reporter mutants.

Figure 5. Summary of fluorescence and current response patterns for the Y89C, Q146C, and

M223C mutants. A-C. Graphical summaries of ligand fluorescence response (ΔFligand) amplitudes normalized to the response amplitude to a previous application of 5-HT (ΔF5-HT) for the Y89C (A),

Q146C (B), and M223C (C) reporter mutants. The 5-HT data represent recordings of a following application of 5-HT following the same standard recording protocol illustrated in Fig.4A. Data

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

represent mean ± S.E.M. for 5 to 16 recordings of individual oocytes. * Significantly different from

5-HT; p < 0.05 (ANOVA with Dunnet’s correction for multiple comparisons).

Figure 6. Concentration-response curves for agonist and antagonist-induced fluorescence changes at the Y89C, Q146C, and M223C mutants. A. Concentration-response curves for fluorescence and current responses evoked by the agonists 5-HT (upper curves) and mCPBG (lower curves). Data points represent the mean ± SEM from 5 to 8 individual experiments. concentration- dependent increases in fluorescence. B. Representative traces illustrating the standard recording Downloaded from protocol employed for fluorescence concentration-response experiments. Shown are 5-second segments from recording traces of membrane current (black traces) and fluorescence (red traces) molpharm.aspetjournals.org from an oocyte expressing TAMRA-conjugated Y89C mutant receptor during application of increasing concentrations of ondansetron (indicated by black bars) (Materials & methods). The excitation LED was pulsed for 200 ms (indicated by green bars) every second to limit the

fluorophore photodestruction during the extended recording time. Each ligand concentration was at ASPET Journals on October 1, 2021 applied for two minutes to allow stable fluorescence levels. C. Concentration-response curves for fluorescence responses evoked by setron-class antagonists at the Y89C, Q146C, and M223C mutants. Data points represent the mean ± SD from 4 to 8 individual experiments.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

13. Tables

Table 1: Residues in h5-HT3A selected for cysteine substitution and their equivalents in other Cys-loop receptor subunits previously studied by fluorescence.

Human h5-HT3A Previously studied positions in Cys-loop receptors as potential VCF Region reporters human ρ1 GABAA Rat α1 GABAA Human α1 GlyR

Asp74# Ser66a N.T. Ala52d Loop 2 Tyr89 N.T. N.T. Gln67 d Loop D Downloaded from Gln146 Leu166a Leu127c Leu127 d Loop E Val196# Asp214b N.T. N.T. Loop F Phe221 N.T. N.T. N.T. Loop C Met223 Tyr241a N.T. His201 d Loop C

Ser225# N.T. N.T. Asn203 d Loop C molpharm.aspetjournals.org Ile238# N.T. N.T. N.T. Pre-M1 Arg239# N.T. N.T. Glu217 d Pre-M1 Arg241# N.T. N.T. Gln219 d Pre-M1 Leu243# N.T. N.T. Gly221 d Pre-M1 Val246# N.T. N.T. Leu224 d Pre-M1 # d

Val247 N.T. N.T. Ile225 Pre-M1 at ASPET Journals on October 1, 2021 Leu249# N.T. N.T. Met227 d Pre-M1 a Chang et al. (2002), b Zhang et al. (2009), c Muroi et al. (2006) and dPless & Lynch (2009a). # Indicates that the position did not report fluorescence changes when mutated to Cys and labeled with a fluorophore in the indicated study. Residue numbering according to original publications and may differ from canonical sequence numbering. Other studies concerning studies of Cys-loop receptors by fluorescence labeling at other positions thank those in the present study include Mourot et al. (2008), Menny et al. (2017), and

Polovinkin et al., (2018). N. T.; not tested in the listed subunit.

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Table 2: EC50 values for current and fluorescence responses to 5-HT and mCPBG at unlabelled and labeled

WT and mutant h5-HT3A receptors.

a a Mutant 5-HT EC50 (µM) mCPGB EC50 (µM)

TEVC ∆F TEVC ∆F

WT 2.6 [2.2;3.0] - 0.45 [0.35;0.60] -

WT + TAMRA 3.5 [3.1;4.7]* - 0.95 [0.70;1.3] - Downloaded from

Y89C 0.24 [0.20;0.28]* - 0.48 [0.34;0.68] -

Y89C + TAMRA 0.70 [0.56;0.86]* 4.0 [3.2 ; 4.9]# 0.26 [0.20;0.34] 1.0 [0.67;1.5] molpharm.aspetjournals.org Q146C 1.2 [0.9;1.7]* - 1.5 [1.4;2.0] -

Q146C + TAMRA 4.8 [3.0;7.9] 61 [48 ; 79]# 1.1 [0.90;1.45] 49 [36;65]#

F221C 85 [76;94]* - - -

F221C + TAMRA >300* >300 - - at ASPET Journals on October 1, 2021 M223C 7.2 [5.2;10]* - 3.7 [ 3.0;4.5]* -

M223C + TAMRA 22 [19;26]* 18 [14 ; 23] 8.9 [6.1;13]* 24 [15;38] a EC50 values were determined as described in Methods and materials by nonlinear fitting of concentration- inhibition data collected at 5 to 10 oocytes. Numbers in bracket denote the 95% confidence interval for EC50.

# *P <0.01 versus WT. P <0.01 versus TEVC EC50 (ANOVA with Dunnet’s correction for multiple comparisons).

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Table 3: EC50 values for fluorescence responses to antagonists at mutant h5-HT3A receptors.

a Reporter mutant EC50 (µM)

Ondansetron Granisetron Encenicline Tropisetron

Y89C 58 [36,96] 100 [59;180] 170 [110;270] 45 [24;87]

Q146C 850 [73;1000] 800 [670;970] 370 [340;410] 850 [540;1300]

M223C 51 [38;70] 54 [38;77] 74 [59;95] 38 [19;74] a EC50 values were determined as described in Methods and materials by non-linear fitting of concentration- Downloaded from inhibition data collected at 5 to 10 oocytes. Numbers in bracket denote the 95% confidence interval for EC50. molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

14. Figures

Figure 1 Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Figure 2 Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Figure 3 Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Figure 4 Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Figure 5 Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021

Molecular Pharmacology Fast Forward. Published on October 3, 2019 as DOI: 10.1124/mol.119.116657 This article has not been copyedited and formatted. The final version may differ from this version.

MOL # 116657

Figure 6 Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 1, 2021